Multi-scale predictive habitat suitability modeling based on hierarchically delineated patches: an example for yellow-billed cuckoos nesting in riparian forests,California, USA

The discipline of landscape ecology recognizes the importance of measuring habitat suitability variables at spatial scales relevant to specific organisms. This paper uses a novel multi-scale hierarchical patch delineation method, PatchMorph, to measure landscape patch characteristics at two distinct spatial scales and statistically relate them to the presence of state-listed endangered yellow-billed cuckoos (Coccyzus americanus occidentalis) nesting in forest patches along the Sacramento River, California, USA. The landscape patch characteristics calculated were: patch thickness, area of cottonwood forest, area of riparian scrub, area of other mixed riparian forest, and total patch area. A third, regional spatial variable, delineating the north and south portions of study area was also analyzed for the effect of regional processes. Using field surveys, the landscape characteristics were related to patch occupancy by yellow-billed cuckoos. The area of cottonwood forest measured at the finest spatial scale of patches was found to be the most important factor determining yellow-billed cuckoo presence in the forest patches, while no patch characteristics at the larger scale of habitat patches were important. The regional spatial variable was important in two of the three analysis techniques. Model validation using an independent data set of surveys (conducted 1987–1990) found 76–82% model accuracy for all the statistical techniques used. Our results show that the spatial scale at which habitat characteristics are measured influences the suitability of forest patches. This multi-scale patch and model selection approach to habitat suitability analysis can readily be generalized for use with other organisms and systems.     

Combining top-down and bottom-up dynamics in land use modeling: exploring the future of abandoned farmlands in Europe with the Dyna-CLUE model

Land use change is the result of interactions between processes operating at different scales. Simulation models at regional to global scales are often incapable of including locally determined processes of land use change. This paper introduces a modeling approach that integrates demand-driven changes in land area with locally determined conversion processes. The model is illustrated with an application for European land use. Interactions between changing demands for agricultural land and vegetation processes leading to the re-growth of (semi-) natural vegetation on abandoned farmland are explicitly addressed. Succession of natural vegetation is simulated based on the spatial variation in biophysical and management related conditions, while the dynamics of the agricultural area are determined by a global multi-sector model. The results allow an exploration of the future dynamics of European land use and landscapes. The model approach is similarly suitable for other regions and processes where large scale processes interact with local dynamics.     

Confronting collinearity: comparing methods for disentangling the effects of habitat loss and fragmentation

Estimating the relative importance of habitat loss and fragmentation is necessary to estimate the potential benefits of specific management actions and to ensure that limited conservation resources are used efficiently. However, estimating relative effects is complicated because the two processes are highly correlated. Previous studies have used a wide variety of statistical methods to separate their effects and we speculated that the published results may have been influenced by the methods used. We used simulations to determine whether, under identical conditions, the following 7 methods generate different estimates of relative importance for realistically correlated landscape predictors: residual regression, model or variable selection, averaged coefficients from all supported models, summed Akaike weights, classical variance partitioning, hierarchical variance partitioning, and a multiple regression model with no adjustments for collinearity. We found that different methods generated different rankings of the predictors and that some metrics were strongly biased. Residual regression and variance partitioning were highly biased by correlations among predictors and the bias depended on the direction of a predictor’s effect (positive vs. negative). Our results suggest that many efforts to deal with the correlation between amount and fragmentation may have done more harm than good. If confounding effects are controlled and adequate thought is given to the ecological mechanisms behind modeled predictors, then standardized partial regression coefficients are unbiased estimates of the relative importance of amount and fragmentation, even when predictors are highly correlated.     

Changes in fox squirrel anti-predator behaviors across the urban–rural gradient

Predator stimuli created by humans in the urban environment may alter animals’ anti-predator behaviors. I hypothesized that habituation would cause anti-predator behaviors to decrease in urban settings in response to humans. Additionally, I hypothesized that populations habituated to humans would show reduced responses to other predator stimuli. I observed three populations of squirrels (urban, suburban and rural) responses to human approaches, red-tailed hawk vocalizations (Buteo jamaicensis) and coyote (Canis latrans) vocalizations. Mahalanobis distances of anti-predator behaviors in response to human approaches were consistent with the urban–rural gradient. Flight initiation distances (X ² = 26.33, df = 2, P < 0.001) and amount of time dedicated to anti-predator behavior (X ² = 10.94, df = 2, P = 0.004) in response to human approaches were also consistent with the urban–rural gradient. Supporting the habituation hypothesis, naive juvenile squirrels increased flight initiation distances (X ² = 35.89, df = 1, P < 0.001) and time dedicated to anti-predator behaviors (X ² = 9.46, df = 1, P = 0.002) relative to adult squirrels in the same urban environment. Time dedicated to anti-predator behaviors differed among all three sites in response to both coyote (X ² = 9.83, df = 2, P = 0.007) and hawk (X ² = 6.50, df = 2, P = 0.035) vocalizations. Responses to both vocalizations on rural sites (coyote = 45%, hawk = 55%) greater than twice that found on the urban sites (coyote = 11%, hawk = 20%). This is possibly the first case of a transfer of habituation demonstrated under field conditions.     

Effects of multiple fires on tree invasion in montane grasslands

There is circumstantial evidence that grasslands on the Bunya Mountains were once maintained by Aboriginal burning, and with lack of fire under European management are being colonised by trees. To assess the efficacy of burning for maintaining grasslands, 119 fires were lit between 1996 and 2006. The total area of unburnt grasslands decreased by 27%, while grasslands burnt at least once decreased by 1%. The density of invading trees was recorded from fixed plots on 23 grasslands burnt between one and six times. Cassinia was virtually eliminated and the density of the Rainforest species guild slowly but continually declined. Acacia irrorata exhibited a humped response, with initial increases resulting from vegetative resprouting and gradual decline with persistent burning. Phyllodinous Acacia and Woodland trees were the least fire sensitive guilds, having stable or increased density with repeated burning. Multi-factor regression modelling detected no significant relationships between changes in woody plant density and the interval between fires, fire intensity, the initial density of large trees, an index of soil moisture, or the cumulative number of fires for any species guild. The survivorship of both Cassinia and Rainforest guilds was significantly lower with summer burning than winter burning, but a seasonal effect of burning was not evident for other guilds. The findings suggest that regardless of fire conditions, frequent burning will reduce the number of adult trees, maintain resprouts in an immature state, facilitate further fire and reduce the rate of grassland loss. Woodland species are especially resilient to fire, and burning to maintain grassy ecosystems will be most successful where the main colonisers are rainforest species and burning is conducted in summer. The findings suggest that the montane grasslands of the Bunya Mountains were maintained by anthropogenic burning and active fire management will prolong their existence.     

Effects of water table management on soil salinity and alfalfa yield in a semi-arid climate

A lysimeter experiment was conducted to investigate the effect of water table management (WTM) on distribution of soil salinity and annual alfalfa (Medicago scutellata) yield. Subirrigations with three levels of water table namely, 0.5 (WT0.5), 0.7 (WT0.7), and 1.0 m (WT1.0) and a free drainage (FD) conventional irrigation treatment were selected for this study. All treatments were arranged in a complete randomized block design with three replicates. The results of this study indicated that the average soil electrical conductivity of the saturated extract (ECe) in the root zone gradually increased and exceeded the designated crop threshold value (4 dS/m) after the first forage harvest in subirrigated lysimeters. A higher salt accumulation was observed at the WT0.5 treatment. The average dry matter yield of annual alfalfa in WT0.5 and WT0.7 treatments was found to be 52 and 73% higher compared with the control treatment, respectively.     

Prediction of irrigation return flows through a hierarchical modeling approach

The quantity of water available for irrigation is getting scarce in many countries and it assumes great importance for assured crop production, especially in view of the erratic behavior of the monsoon. Thus, there is a pressing need to improve the water efficiency of irrigation systems. One-way of improving the efficiency of the irrigation system is reusing the return flow from the irrigation system. This task requires quantification of return flow, which still remains as a grey area in irrigation water management. The estimation of return flow from the irrigation system is usually obtained using thumb rules depending upon the site-specific conditions like command area conditions and soil properties. In this paper, a hierarchical modeling technique, namely, regression tree is developed for return flow estimation. Regression tree is built through binary recursive partitioning. The effective rainfall, inflow, consumptive water demand, and percolation loss are taken as predictor variables and return flow is treated as the target variable. The applicability of the hierarchical model is demonstrated through a case study of Periyar-Vaigai Irrigation System in Tamil Nadu, India. The model performance shows a good match between the simulated and the field measured return flow values. Results of statistical analysis indicated that the correlation coefficients are high for both single as well as double crop seasons.     

Bahiagrass crop coefficients from eddy correlation measurements in central Florida

Bahiagrass (Paspalum notatum) is a warm-season grass used primarily in pastures and along highways and other low maintenance public areas in Florida. It is also used in landscapes to some extent because of its drought tolerance. Bahiagrass can survive under a range of moisture conditions from no irrigation to very wet conditions. Its well-watered consumptive use has not been reported previously. In this study, bahiagrass crop coefficients (K _c) for an irrigated pasture were determined for July 2003 through December 2006 in central Florida. The eddy correlation method was used to estimate crop evapotranspiration (ET_c) rates. The standardized reference evapotranspiration (ET_o) equation (ASCE-EWRI standardization of reference evapotranspiration task committee report, 2005) was applied to calculate ET_o values using on site weather data. Daily K _c values were estimated from the ratio of the measured ET_c and the calculated ET_o. The recommended K _c values for bahiagrass are 0.35 for January–February, 0.55 for March, 0.80 for April, 0.90 for May, 0.75 for June, 0.70 for July–August, 0.75 for September, 0.70 for October, 0.60 for November, and 0.45 for December in central Florida. The highest K _c value of 0.9 in May corresponded with maximum vapor pressure deficit conditions as well as cloud free conditions and the highest incoming solar radiation as compared to the rest of the year. During the summer (June to August), frequent precipitation events increased the cloud cover and reduced grass water use. The K _c annual trend was similar to estimated K _c values from another well-watered warm-season grass study in Florida.     

Evaluating eddy covariance cotton ET measurements in an advective environment with large weighing lysimeters

Eddy covariance (EC) systems are being used to assess the accuracy of remote sensing methods in mapping surface sensible and latent heat fluxes and evapotranspiration (ET) from local to regional scales, and in crop coefficient development. Therefore, the objective was to evaluate the accuracy of EC systems in measuring sensible heat (H) and latent heat (LE) fluxes. For this purpose, two EC systems were installed near large monolithic weighing lysimeters, on irrigated cotton fields in the Texas High Plains, during the months of June and July 2008. Sensible and latent heat fluxes were underestimated with an average error of about 30%. Most of the errors were from nocturnal measurements. Energy balance (EB) closure was 73.2–78.0% for daytime fluxes. Thus, daylight fluxes were adjusted for lack of EB closure using the Bowen ratio/preservation of energy principle, which improved the resulting EC heat flux agreement with lysimetric values. Further adjustments to EC-based ET included nighttime ET (composite) incorporation, and the use of ‘heat flux source area’ (footprint) functions to compensate ET when the footprint expanded beyond the crop field boundary. As a result, ET values remarkably matched lysimetric ET values, with a ‘mean bias error ± root mean square error’ of −0.03 ± 0.5 mm day⁻¹ (or −0.6 ± 10.2%).